Convolutional Neural Networks

Project: Write an Algorithm for a Dog Identification App


In this notebook, some template code has already been provided for you, and you will need to implement additional functionality to successfully complete this project. You will not need to modify the included code beyond what is requested. Sections that begin with '(IMPLEMENTATION)' in the header indicate that the following block of code will require additional functionality which you must provide. Instructions will be provided for each section, and the specifics of the implementation are marked in the code block with a 'TODO' statement. Please be sure to read the instructions carefully!

Note: Once you have completed all of the code implementations, you need to finalize your work by exporting the Jupyter Notebook as an HTML document. Before exporting the notebook to html, all of the code cells need to have been run so that reviewers can see the final implementation and output. You can then export the notebook by using the menu above and navigating to File -> Download as -> HTML (.html). Include the finished document along with this notebook as your submission.

In addition to implementing code, there will be questions that you must answer which relate to the project and your implementation. Each section where you will answer a question is preceded by a 'Question X' header. Carefully read each question and provide thorough answers in the following text boxes that begin with 'Answer:'. Your project submission will be evaluated based on your answers to each of the questions and the implementation you provide.

Note: Code and Markdown cells can be executed using the Shift + Enter keyboard shortcut. Markdown cells can be edited by double-clicking the cell to enter edit mode.

The rubric contains optional "Stand Out Suggestions" for enhancing the project beyond the minimum requirements. If you decide to pursue the "Stand Out Suggestions", you should include the code in this Jupyter notebook.


Why We're Here

In this notebook, you will make the first steps towards developing an algorithm that could be used as part of a mobile or web app. At the end of this project, your code will accept any user-supplied image as input. If a dog is detected in the image, it will provide an estimate of the dog's breed. If a human is detected, it will provide an estimate of the dog breed that is most resembling. The image below displays potential sample output of your finished project (... but we expect that each student's algorithm will behave differently!).

Sample Dog Output

In this real-world setting, you will need to piece together a series of models to perform different tasks; for instance, the algorithm that detects humans in an image will be different from the CNN that infers dog breed. There are many points of possible failure, and no perfect algorithm exists. Your imperfect solution will nonetheless create a fun user experience!

The Road Ahead

We break the notebook into separate steps. Feel free to use the links below to navigate the notebook.

  • Step 0: Import Datasets
  • Step 1: Detect Humans
  • Step 2: Detect Dogs
  • Step 3: Create a CNN to Classify Dog Breeds (from Scratch)
  • Step 4: Create a CNN to Classify Dog Breeds (using Transfer Learning)
  • Step 5: Write your Algorithm
  • Step 6: Test Your Algorithm

Step 0: Import Datasets

Make sure that you've downloaded the required human and dog datasets:

Note: if you are using the Udacity workspace, you DO NOT need to re-download these - they can be found in the /data folder as noted in the cell below.

  • Download the dog dataset. Unzip the folder and place it in this project's home directory, at the location /dog_images.

  • Download the human dataset. Unzip the folder and place it in the home directory, at location /lfw.

Note: If you are using a Windows machine, you are encouraged to use 7zip to extract the folder.

In the code cell below, we save the file paths for both the human (LFW) dataset and dog dataset in the numpy arrays human_files and dog_files.

In [1]:
#!wget -qq https://s3-us-west-1.amazonaws.com/udacity-aind/dog-project/dogImages.zip
#!unzip -qq dogImages.zip
#!rm dogImages.zip
#!wget -qq https://s3-us-west-1.amazonaws.com/udacity-aind/dog-project/lfw.zip
#!unzip -qq lfw.zip
#!rm lfw.zip
In [2]:
import numpy as np
from glob import glob

# load filenames for human and dog images
human_files = np.array(glob("/data/lfw/*/*"))
dog_files = np.array(glob("/data/dog_images/*/*/*"))

# print number of images in each dataset
print('There are %d total human images.' % len(human_files))
print('There are %d total dog images.' % len(dog_files))
There are 13233 total human images.
There are 8351 total dog images.

Step 1: Detect Humans

In this section, we use OpenCV's implementation of Haar feature-based cascade classifiers to detect human faces in images.

OpenCV provides many pre-trained face detectors, stored as XML files on github. We have downloaded one of these detectors and stored it in the haarcascades directory. In the next code cell, we demonstrate how to use this detector to find human faces in a sample image.

In [3]:
import cv2                
import matplotlib.pyplot as plt                        
%matplotlib inline                               

# extract pre-trained face detector
face_cascade = cv2.CascadeClassifier('haarcascades/haarcascade_frontalface_alt.xml')

# load color (BGR) image
img = cv2.imread(human_files[0])
# convert BGR image to grayscale
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)

# find faces in image
faces = face_cascade.detectMultiScale(gray)

# print number of faces detected in the image
print('Number of faces detected:', len(faces))

# get bounding box for each detected face
for (x,y,w,h) in faces:
    # add bounding box to color image
    cv2.rectangle(img,(x,y),(x+w,y+h),(255,0,0),2)
    
# convert BGR image to RGB for plotting
cv_rgb = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)

# display the image, along with bounding box
plt.imshow(cv_rgb)
plt.show()
Number of faces detected: 1

Before using any of the face detectors, it is standard procedure to convert the images to grayscale. The detectMultiScale function executes the classifier stored in face_cascade and takes the grayscale image as a parameter.

In the above code, faces is a numpy array of detected faces, where each row corresponds to a detected face. Each detected face is a 1D array with four entries that specifies the bounding box of the detected face. The first two entries in the array (extracted in the above code as x and y) specify the horizontal and vertical positions of the top left corner of the bounding box. The last two entries in the array (extracted here as w and h) specify the width and height of the box.

Write a Human Face Detector

We can use this procedure to write a function that returns True if a human face is detected in an image and False otherwise. This function, aptly named face_detector, takes a string-valued file path to an image as input and appears in the code block below.

In [4]:
# returns "True" if face is detected in image stored at img_path
def face_detector(img_path):
    img = cv2.imread(img_path)
    gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
    faces = face_cascade.detectMultiScale(gray)
    return len(faces) > 0

(IMPLEMENTATION) Assess the Human Face Detector

Question 1: Use the code cell below to test the performance of the face_detector function.

  • What percentage of the first 100 images in human_files have a detected human face?
  • What percentage of the first 100 images in dog_files have a detected human face?

Ideally, we would like 100% of human images with a detected face and 0% of dog images with a detected face. You will see that our algorithm falls short of this goal, but still gives acceptable performance. We extract the file paths for the first 100 images from each of the datasets and store them in the numpy arrays human_files_short and dog_files_short.

Answer:
Human faces detected 98.0 %, and not detected in 56 and 87 images.
Dog images detected in human face classifier are 17%, but I do think it is bias result, but the human face detector detected 17%

In [5]:
from tqdm import tqdm

human_files_short = human_files[:100]
dog_files_short = dog_files[:100]

#-#-# Do NOT modify the code above this line. #-#-#

## TODO: Test the performance of the face_detector algorithm 
count = 0
not_found = []
for i in range(len(human_files_short)):
    face_detect = face_detector(human_files_short[i])
    if face_detect == True:
        count+=1
    else:
        not_found.append(i)
        
count = count/100 * 100
print("Human faces detected",count,'%')
print("Human faces not detected in",not_found)
## on the images in human_files_short and dog_files_short.
Human faces detected 98.0 %
Human faces not detected in [56, 87]
In [6]:
count = 0
not_found = []
for i in range(len(dog_files_short)):
    face_detect = face_detector(dog_files_short[i])
    if face_detect == True:
        count+=1
    else:
        not_found.append(i)
        
count = count/100 * 100
print("Dog faces detected",count,'%')
print("Dog faces not detected in",not_found)
## on the images in human_files_short and dog_files_short.
Dog faces detected 17.0 %
Dog faces not detected in [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 41, 43, 44, 45, 46, 47, 48, 50, 51, 52, 53, 54, 55, 57, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 72, 73, 74, 75, 76, 78, 80, 81, 82, 83, 85, 86, 87, 91, 92, 93, 94, 96, 97, 98, 99]

We suggest the face detector from OpenCV as a potential way to detect human images in your algorithm, but you are free to explore other approaches, especially approaches that make use of deep learning :). Please use the code cell below to design and test your own face detection algorithm. If you decide to pursue this optional task, report performance on human_files_short and dog_files_short.

In [7]:
### (Optional) 
### TODO: Test performance of anotherface detection algorithm.
### Feel free to use as many code cells as needed.

Step 2: Detect Dogs

In this section, we use a pre-trained model to detect dogs in images.

Obtain Pre-trained VGG-16 Model

The code cell below downloads the VGG-16 model, along with weights that have been trained on ImageNet, a very large, very popular dataset used for image classification and other vision tasks. ImageNet contains over 10 million URLs, each linking to an image containing an object from one of 1000 categories.

In [8]:
import torch
import torchvision.models as models

# define VGG16 model
VGG16 = models.vgg16(pretrained=True)

# check if CUDA is available
use_cuda = torch.cuda.is_available()

# move model to GPU if CUDA is available
if use_cuda:
    VGG16 = VGG16.cuda()
Downloading: "https://download.pytorch.org/models/vgg16-397923af.pth" to /root/.torch/models/vgg16-397923af.pth
100%|██████████| 553433881/553433881 [00:05<00:00, 100522506.45it/s]

Given an image, this pre-trained VGG-16 model returns a prediction (derived from the 1000 possible categories in ImageNet) for the object that is contained in the image.

(IMPLEMENTATION) Making Predictions with a Pre-trained Model

In the next code cell, you will write a function that accepts a path to an image (such as 'dogImages/train/001.Affenpinscher/Affenpinscher_00001.jpg') as input and returns the index corresponding to the ImageNet class that is predicted by the pre-trained VGG-16 model. The output should always be an integer between 0 and 999, inclusive.

Before writing the function, make sure that you take the time to learn how to appropriately pre-process tensors for pre-trained models in the PyTorch documentation.

In [9]:
from PIL import Image
import torchvision.transforms as transforms

def VGG16_predict(img_path):
    '''
    Use pre-trained VGG-16 model to obtain index corresponding to 
    predicted ImageNet class for image at specified path
    
    Args:
        img_path: path to an image
        
    Returns:
        Index corresponding to VGG-16 model's prediction
    '''
    
    ## TODO: Complete the function.
    ## Load and pre-process an image from the given img_path
    
    transform = transforms.Compose([transforms.Resize(256),transforms.CenterCrop(224),
                                         transforms.ToTensor(),
                                         transforms.Normalize(mean=[0.485, 0.456, 0.406],
                                                              std=[0.229, 0.224, 0.225])])
    img = Image.open(img_path)
    transform_image = transform(img)
    squ = torch.unsqueeze(transform_image, 0)
    #img = Variable(squ)
    VGG16.eval()
    
    ## Load and pre-process an image from the given img_path
    ## Return the *index* of the predicted class for that image
    prediction = VGG16(squ.cuda())  # Returns a Tensor of shape (batch, num class labels)
    prediction = prediction.cpu().data.numpy().argmax()
   # prediction = prediction.data.numpy().argmax()
     # predicted class index
    ## Return the *index* of the predicted class for that image
    return prediction
     # predicted class index

(IMPLEMENTATION) Write a Dog Detector

While looking at the dictionary, you will notice that the categories corresponding to dogs appear in an uninterrupted sequence and correspond to dictionary keys 151-268, inclusive, to include all categories from 'Chihuahua' to 'Mexican hairless'. Thus, in order to check to see if an image is predicted to contain a dog by the pre-trained VGG-16 model, we need only check if the pre-trained model predicts an index between 151 and 268 (inclusive).

Use these ideas to complete the dog_detector function below, which returns True if a dog is detected in an image (and False if not).

In [10]:
### returns "True" if a dog is detected in the image stored at img_path
def dog_detector(img_path):
    ## TODO: Complete the function.
    ## TODO: Complete the function.
    pred = VGG16_predict(img_path)
    #i tried below for intiall dog detection, but gets wrong result
    #prediction = pred > 0
    result = (True if pred in range(151,268) else False)
    return result # true/false
    #return None # true/false
In [11]:
img_path = 'dogImages/train/001.Affenpinscher/Affenpinscher_00001.jpg'
dog_detector(img_path)
Out[11]:
True

(IMPLEMENTATION) Assess the Dog Detector

Question 2: Use the code cell below to test the performance of your dog_detector function.

  • What percentage of the images in human_files_short have a detected dog?
  • What percentage of the images in dog_files_short have a detected dog?

Answer:
Dog faces detected 100.0 %

Human faces detected in dog detector is 0.0%

In [12]:
### TODO: Test the performance of the dog_detector function
count = 0
not_found = []
for i in range(len(dog_files_short)):
    face_detect = dog_detector(dog_files_short[i])
    if face_detect == True:
        count+=1
    else:
        not_found.append(i)
        
count = count/100 * 100
print("Dog faces detected",count,'%')
print("Dog faces not detected in",not_found)
### on the images in human_files_short and dog_files_short.
Dog faces detected 100.0 %
Dog faces not detected in []
In [13]:
count = 0
not_found = []
for i in range(len(human_files_short)):
    face_detect = dog_detector(human_files_short[i])
    if face_detect == True:
        count+=1
    else:
        not_found.append(i)
        
count = count/100 * 100
print("Human faces detected",count,'%')
print("Human faces not detected in",not_found)
Human faces detected 0.0 %
Human faces not detected in [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99]

We suggest VGG-16 as a potential network to detect dog images in your algorithm, but you are free to explore other pre-trained networks (such as Inception-v3, ResNet-50, etc). Please use the code cell below to test other pre-trained PyTorch models. If you decide to pursue this optional task, report performance on human_files_short and dog_files_short.

In [14]:
### (Optional) 
### TODO: Report the performance of another pre-trained network.
### Feel free to use as many code cells as needed.

Step 3: Create a CNN to Classify Dog Breeds (from Scratch)

Now that we have functions for detecting humans and dogs in images, we need a way to predict breed from images. In this step, you will create a CNN that classifies dog breeds. You must create your CNN from scratch (so, you can't use transfer learning yet!), and you must attain a test accuracy of at least 10%. In Step 4 of this notebook, you will have the opportunity to use transfer learning to create a CNN that attains greatly improved accuracy.

We mention that the task of assigning breed to dogs from images is considered exceptionally challenging. To see why, consider that even a human would have trouble distinguishing between a Brittany and a Welsh Springer Spaniel.

Brittany Welsh Springer Spaniel

It is not difficult to find other dog breed pairs with minimal inter-class variation (for instance, Curly-Coated Retrievers and American Water Spaniels).

Curly-Coated Retriever American Water Spaniel

Likewise, recall that labradors come in yellow, chocolate, and black. Your vision-based algorithm will have to conquer this high intra-class variation to determine how to classify all of these different shades as the same breed.

Yellow Labrador Chocolate Labrador Black Labrador

We also mention that random chance presents an exceptionally low bar: setting aside the fact that the classes are slightly imabalanced, a random guess will provide a correct answer roughly 1 in 133 times, which corresponds to an accuracy of less than 1%.

Remember that the practice is far ahead of the theory in deep learning. Experiment with many different architectures, and trust your intuition. And, of course, have fun!

(IMPLEMENTATION) Specify Data Loaders for the Dog Dataset

Use the code cell below to write three separate data loaders for the training, validation, and test datasets of dog images (located at dog_images/train, dog_images/valid, and dog_images/test, respectively). You may find this documentation on custom datasets to be a useful resource. If you are interested in augmenting your training and/or validation data, check out the wide variety of transforms!

In [15]:
import torch
use_cuda = torch.cuda.is_available()
In [16]:
import os
from torchvision import datasets
from PIL import ImageFile
import os
import torch
from torchvision import datasets, transforms
ImageFile.LOAD_TRUNCATED_IMAGES = True

### TODO: Write data loaders for training, validation, and test sets
transforms = {
       'train': transforms.Compose([
        transforms.Resize(256),
        transforms.CenterCrop(224),
        transforms.RandomHorizontalFlip(),
        transforms.ToTensor(),
        transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])
    ]),
    'valid' : transforms.Compose([transforms.Resize(256),
                                  transforms.CenterCrop(224),
                                  transforms.ToTensor(),
                                  transforms.Normalize(mean=[0.485, 0.456, 0.406],
                                                       std=[0.229, 0.224, 0.225])]),
    
    'test' : transforms.Compose([transforms.Resize(256),
                                 transforms.CenterCrop(224),
                                 transforms.ToTensor(),
                                 transforms.Normalize(mean=[0.485, 0.456, 0.406],
                                                  std=[0.229, 0.224, 0.225])])
}
num_workers = 0
batch_sizes = 20
## Specify appropriate transforms, and batch_sizes
image_datasets = {x: datasets.ImageFolder(os.path.join("/data/dog_images/", x), transforms[x])
                 for x in ['train', 'valid', 'test']}

loaders_scratch = {x: torch.utils.data.DataLoader(image_datasets[x], batch_size=batch_sizes,
                                              shuffle=True, num_workers=num_workers)
               for x in ['train', 'valid', 'test']}
In [ ]:
 
In [17]:
#displaying images from train folder
def imshow(inp, title=None):
    """Imshow for Tensor."""
    inp = inp.numpy().transpose((1, 2, 0))
    mean = np.array([0.485, 0.456, 0.406])
    std = np.array([0.229, 0.224, 0.225])
    inp = std * inp + mean
    inp = np.clip(inp, 0, 1)
    plt.imshow(inp)
    if title is not None:
        plt.title(title)
    plt.pause(0.001) 
In [18]:
import matplotlib.pyplot as plt
import numpy as np
images, classes = next(iter(loaders_scratch['train']))

# Make a grid from batch
fig = plt.figure(figsize=(25, 4))
# display 20 images
for idx in np.arange(20):
    ax = fig.add_subplot(2, 20/2, idx+1, xticks=[], yticks=[])
    imshow(images[idx])
    #ax.set_title(dog_label[idx])
In [19]:
dog_label = image_datasets['train'].classes
print(dog_label)
print(len(dog_label))
print(images.shape)
['001.Affenpinscher', '002.Afghan_hound', '003.Airedale_terrier', '004.Akita', '005.Alaskan_malamute', '006.American_eskimo_dog', '007.American_foxhound', '008.American_staffordshire_terrier', '009.American_water_spaniel', '010.Anatolian_shepherd_dog', '011.Australian_cattle_dog', '012.Australian_shepherd', '013.Australian_terrier', '014.Basenji', '015.Basset_hound', '016.Beagle', '017.Bearded_collie', '018.Beauceron', '019.Bedlington_terrier', '020.Belgian_malinois', '021.Belgian_sheepdog', '022.Belgian_tervuren', '023.Bernese_mountain_dog', '024.Bichon_frise', '025.Black_and_tan_coonhound', '026.Black_russian_terrier', '027.Bloodhound', '028.Bluetick_coonhound', '029.Border_collie', '030.Border_terrier', '031.Borzoi', '032.Boston_terrier', '033.Bouvier_des_flandres', '034.Boxer', '035.Boykin_spaniel', '036.Briard', '037.Brittany', '038.Brussels_griffon', '039.Bull_terrier', '040.Bulldog', '041.Bullmastiff', '042.Cairn_terrier', '043.Canaan_dog', '044.Cane_corso', '045.Cardigan_welsh_corgi', '046.Cavalier_king_charles_spaniel', '047.Chesapeake_bay_retriever', '048.Chihuahua', '049.Chinese_crested', '050.Chinese_shar-pei', '051.Chow_chow', '052.Clumber_spaniel', '053.Cocker_spaniel', '054.Collie', '055.Curly-coated_retriever', '056.Dachshund', '057.Dalmatian', '058.Dandie_dinmont_terrier', '059.Doberman_pinscher', '060.Dogue_de_bordeaux', '061.English_cocker_spaniel', '062.English_setter', '063.English_springer_spaniel', '064.English_toy_spaniel', '065.Entlebucher_mountain_dog', '066.Field_spaniel', '067.Finnish_spitz', '068.Flat-coated_retriever', '069.French_bulldog', '070.German_pinscher', '071.German_shepherd_dog', '072.German_shorthaired_pointer', '073.German_wirehaired_pointer', '074.Giant_schnauzer', '075.Glen_of_imaal_terrier', '076.Golden_retriever', '077.Gordon_setter', '078.Great_dane', '079.Great_pyrenees', '080.Greater_swiss_mountain_dog', '081.Greyhound', '082.Havanese', '083.Ibizan_hound', '084.Icelandic_sheepdog', '085.Irish_red_and_white_setter', '086.Irish_setter', '087.Irish_terrier', '088.Irish_water_spaniel', '089.Irish_wolfhound', '090.Italian_greyhound', '091.Japanese_chin', '092.Keeshond', '093.Kerry_blue_terrier', '094.Komondor', '095.Kuvasz', '096.Labrador_retriever', '097.Lakeland_terrier', '098.Leonberger', '099.Lhasa_apso', '100.Lowchen', '101.Maltese', '102.Manchester_terrier', '103.Mastiff', '104.Miniature_schnauzer', '105.Neapolitan_mastiff', '106.Newfoundland', '107.Norfolk_terrier', '108.Norwegian_buhund', '109.Norwegian_elkhound', '110.Norwegian_lundehund', '111.Norwich_terrier', '112.Nova_scotia_duck_tolling_retriever', '113.Old_english_sheepdog', '114.Otterhound', '115.Papillon', '116.Parson_russell_terrier', '117.Pekingese', '118.Pembroke_welsh_corgi', '119.Petit_basset_griffon_vendeen', '120.Pharaoh_hound', '121.Plott', '122.Pointer', '123.Pomeranian', '124.Poodle', '125.Portuguese_water_dog', '126.Saint_bernard', '127.Silky_terrier', '128.Smooth_fox_terrier', '129.Tibetan_mastiff', '130.Welsh_springer_spaniel', '131.Wirehaired_pointing_griffon', '132.Xoloitzcuintli', '133.Yorkshire_terrier']
133
torch.Size([20, 3, 224, 224])

Question 3: Describe your chosen procedure for preprocessing the data.

  • How does your code resize the images (by cropping, stretching, etc)? What size did you pick for the input tensor, and why?

  • Did you decide to augment the dataset? If so, how (through translations, flips, rotations, etc)? If not, why not?

In [ ]:
 

Answer:
In this project, images are resize into 256 pixels and centercrop is 224 and my input tensor is 224, because pixels are taked to center to 224 pixels

yes i used augment the dataset, because get more accuracy in results, and i used random RandomHorizontal Flip

(IMPLEMENTATION) Model Architecture

Create a CNN to classify dog breed. Use the template in the code cell below.

In [23]:
import torch.nn as nn
import torch.nn.functional as F

# define the CNN architecture
class Net(nn.Module):
    ### TODO: choose an architecture, and complete the class
    def __init__(self):
        super(Net, self).__init__()
        ## Define layers of a CNN
        self.conv1 = nn.Conv2d(3,32,3,padding=1)
        #self.bn1 = nn.BatchNorm2d(16)
        self.pool = nn.MaxPool2d(2,2)
        self.conv2 = nn.Conv2d(32,64,3,padding = 1)
        #self.pool2 = nn.MaxPool2d(4,4)
        #self.bn2 = nn.BatchNorm2d(32)
        self.conv3 = nn.Conv2d(64,128,3,padding = 1)
        
        self.conv4 = nn.Conv2d(128,256,3,padding = 1)
        self.conv5 = nn.Conv2d(256,512,3,padding = 1)
        #self.conv4 = nn.Conv2d(64,128,3,padding = 1)
        #self.conv5 = nn.conv2d(128,256,3, padding =1)
        #self.pool3 = nn.MaxPool2d(8,8)
        self.fc1 = nn.Linear(512*7*7,512)
        
        self.fc2 = nn.Linear(512,256)
        self.fc3 = nn.Linear(256,133)
        self.dropout = nn.Dropout(0.25)
    def forward(self, x):
        ## Define forward behavior
        x = self.pool(F.relu(self.conv1(x)))
        x = self.pool(F.relu(self.conv2(x)))
        x = self.pool(F.relu(self.conv3(x)))
        x = self.pool(F.relu(self.conv4(x)))
        x = self.pool(F.relu(self.conv5(x)))
        #print(x.shape)
        #print(x.shape)
        #x = self.pool(F.relu(self.conv4(x)))
        #x = self.pool(F.relu(self.conv5(x)))
        x = x.view(-1,512*7*7)
        x = self.dropout(x)
        x = F.relu(self.fc1(x))
        x = self.dropout(x)
        x = F.relu(self.fc2(x))
        x = self.dropout(x)
        x = self.fc3(x)
        return x
    

#-#-# You so NOT have to modify the code below this line. #-#-#


    # create a new model with these weights
    



model_scratch = Net()
def init_weights(m):
        if type(m) == nn.Linear:
            torch.nn.init.xavier_uniform(m.weight)
            m.bias.data.fill_(0.01)
  
model_scratch.apply(init_weights)
print(model_scratch)
# move tensors to GPU if CUDA is available
if use_cuda:
    model_scratch.cuda()
Net(
  (conv1): Conv2d(3, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  (pool): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
  (conv2): Conv2d(32, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  (conv3): Conv2d(64, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  (conv4): Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  (conv5): Conv2d(256, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
  (fc1): Linear(in_features=25088, out_features=512, bias=True)
  (fc2): Linear(in_features=512, out_features=256, bias=True)
  (fc3): Linear(in_features=256, out_features=133, bias=True)
  (dropout): Dropout(p=0.25)
)
/opt/conda/lib/python3.6/site-packages/ipykernel_launcher.py:60: UserWarning: nn.init.xavier_uniform is now deprecated in favor of nn.init.xavier_uniform_.
In [ ]:
 

Question 4: Outline the steps you took to get to your final CNN architecture and your reasoning at each step.

Answer: In my network, there are five layers of CNN architecture, self.conv1 = nn.Conv2d(3,32,3,padding=1) in the first layer network is images depth is 3, I choose 32 filter channel, 3 3 kernel, padding is 1, the stride is by default 1, Maxpooling is 2 2 kernel and stride is 2,
self.conv2 = nn.Conv2d(32,64,3,padding = 1) in the second layer network is images depth is 32, I choose 64 filter channel, 3 3 kernel, padding is 1, the stride is by default 1, Maxpooling is 2 2 kernel and stride is 2, self.conv3 = nn.Conv2d(64,128,3,padding = 1) in the third layer network is images depth is 64, I choose 128 filter channel, 3 3 kernel, padding is 1, the stride is by default 1, Maxpooling is 2 2 kernel and stride is 2, self.conv4 = nn.Conv2d(128,256,3,padding = 1) in the fourth layer network is images depth is 128, I choose 256 filter channel, 3 3 kernel, padding is 1, the stride is by default 1, Maxpooling is 2 2 kernel and stride is 2, self.conv5 = nn.Conv2d(256,512,3,padding = 1) n the fifth layer network is images depth is 256, I choose 512 filter channel, 3 3 kernel, padding is 1, the stride is by default 1, Maxpooling is 2 2 kernel and stride is 2 and connected to fully connected layer with droupout, and initialized weights technies is also used

(IMPLEMENTATION) Specify Loss Function and Optimizer

Use the next code cell to specify a loss function and optimizer. Save the chosen loss function as criterion_scratch, and the optimizer as optimizer_scratch below.

In [24]:
import torch.optim as optim

### TODO: select loss function
criterion_scratch = nn.CrossEntropyLoss()

### TODO: select optimizer
optimizer_scratch = optim.SGD(model_scratch.parameters(), lr=0.099)
In [ ]:
 

(IMPLEMENTATION) Train and Validate the Model

Train and validate your model in the code cell below. Save the final model parameters at filepath 'model_scratch.pt'.

In [26]:
# the following import is required for training to be robust to truncated images
from PIL import ImageFile
ImageFile.LOAD_TRUNCATED_IMAGES = True

def train(n_epochs, loaders, model, optimizer, criterion, use_cuda, save_path):
    """returns trained model"""
    # initialize tracker for minimum validation loss
    valid_loss_min = np.Inf 
    
    for epoch in range(1, n_epochs+1):
        # initialize variables to monitor training and validation loss
        train_loss = 0.0
        valid_loss = 0.0
        running_corrects=0.0
        ###################
        # train the model #
        ###################
        model.train()
        for batch_idx, (data, target) in enumerate(loaders['train']):
            # move to GPU
            if use_cuda:
                data, target = data.cuda(), target.cuda()
                optimizer.zero_grad()
            # forward pass
            output = model(data)
            # calculate batch loss
            loss = criterion(output, target)
            # backward pass
            loss.backward()
            # parameter update
            optimizer.step()
            # update training loss
            train_loss += loss.item() * data.size(0)
        ######################    
        # validate the model #
        ######################
        model.eval()
        for batch_idx, (data, target) in enumerate(loaders['valid']):
            # move to GPU
            if use_cuda:
                data, target = data.cuda(), target.cuda()
                #optimizer.zero_grad()
            ## update the average validation loss
            with torch.no_grad():
            
                output = model(data)
            # batch loss
            loss = criterion(output, target)
            # update validation loss
            valid_loss += loss.item()*data.size(0)
            _, preds = torch.max(output, 1)
            running_corrects += torch.sum(preds == target.data)
            
        # calculate average losses
        train_loss = train_loss/len(loaders['train'].dataset)
        valid_loss = valid_loss/len(loaders['valid'].dataset)
            
        # print training/validation statistics 
        print("Running corrects",running_corrects)
        print('Epoch: {} \tTraining Loss: {:.6f} \tValidation Loss: {:.6f}'.format(
            epoch, 
            train_loss,
            valid_loss
            ))
        
        ## TODO: save the model if validation loss has decreased
        if valid_loss <= valid_loss_min:
            print('Validation loss decreased ({:.6f} --> {:.6f}).  Saving model ...'.format(
            valid_loss_min,
            valid_loss))
            torch.save(model.state_dict(), save_path)
            valid_loss_min = valid_loss
            
    # return trained model
    return model


# train the model
model_scratch = train(60, loaders_scratch, model_scratch, optimizer_scratch, 
                      criterion_scratch, use_cuda, 'model_scratch.pt')

# load the model that got the best validation accuracy
model_scratch.load_state_dict(torch.load('model_scratch.pt'))
Running corrects tensor(25, device='cuda:0')
Epoch: 1 	Training Loss: 4.739157 	Validation Loss: 4.649334
Validation loss decreased (inf --> 4.649334).  Saving model ...
Running corrects tensor(24, device='cuda:0')
Epoch: 2 	Training Loss: 4.609497 	Validation Loss: 4.515941
Validation loss decreased (4.649334 --> 4.515941).  Saving model ...
Running corrects tensor(32, device='cuda:0')
Epoch: 3 	Training Loss: 4.467443 	Validation Loss: 4.390891
Validation loss decreased (4.515941 --> 4.390891).  Saving model ...
Running corrects tensor(36, device='cuda:0')
Epoch: 4 	Training Loss: 4.323246 	Validation Loss: 4.262729
Validation loss decreased (4.390891 --> 4.262729).  Saving model ...
Running corrects tensor(47, device='cuda:0')
Epoch: 5 	Training Loss: 4.259690 	Validation Loss: 4.264628
Running corrects tensor(44, device='cuda:0')
Epoch: 6 	Training Loss: 4.108203 	Validation Loss: 4.122768
Validation loss decreased (4.262729 --> 4.122768).  Saving model ...
Running corrects tensor(59, device='cuda:0')
Epoch: 7 	Training Loss: 3.979103 	Validation Loss: 4.011533
Validation loss decreased (4.122768 --> 4.011533).  Saving model ...
Running corrects tensor(75, device='cuda:0')
Epoch: 8 	Training Loss: 3.853012 	Validation Loss: 3.958155
Validation loss decreased (4.011533 --> 3.958155).  Saving model ...
Running corrects tensor(72, device='cuda:0')
Epoch: 9 	Training Loss: 3.708003 	Validation Loss: 3.869422
Validation loss decreased (3.958155 --> 3.869422).  Saving model ...
Running corrects tensor(92, device='cuda:0')
Epoch: 10 	Training Loss: 3.565812 	Validation Loss: 3.819518
Validation loss decreased (3.869422 --> 3.819518).  Saving model ...
Running corrects tensor(80, device='cuda:0')
Epoch: 11 	Training Loss: 3.387735 	Validation Loss: 4.078852
Running corrects tensor(97, device='cuda:0')
Epoch: 12 	Training Loss: 3.262266 	Validation Loss: 3.773093
Validation loss decreased (3.819518 --> 3.773093).  Saving model ...
Running corrects tensor(115, device='cuda:0')
Epoch: 13 	Training Loss: 3.050174 	Validation Loss: 3.906928
Running corrects tensor(107, device='cuda:0')
Epoch: 14 	Training Loss: 2.908267 	Validation Loss: 3.894550
Running corrects tensor(99, device='cuda:0')
Epoch: 15 	Training Loss: 2.750506 	Validation Loss: 3.764683
Validation loss decreased (3.773093 --> 3.764683).  Saving model ...
Running corrects tensor(110, device='cuda:0')
Epoch: 16 	Training Loss: 2.601144 	Validation Loss: 3.756410
Validation loss decreased (3.764683 --> 3.756410).  Saving model ...
Running corrects tensor(119, device='cuda:0')
Epoch: 17 	Training Loss: 2.460501 	Validation Loss: 3.817020
Running corrects tensor(140, device='cuda:0')
Epoch: 18 	Training Loss: 2.234755 	Validation Loss: 3.677898
Validation loss decreased (3.756410 --> 3.677898).  Saving model ...
Running corrects tensor(133, device='cuda:0')
Epoch: 19 	Training Loss: 2.125084 	Validation Loss: 3.878174
Running corrects tensor(105, device='cuda:0')
Epoch: 20 	Training Loss: 2.053368 	Validation Loss: 3.890153
Running corrects tensor(119, device='cuda:0')
Epoch: 21 	Training Loss: 1.933928 	Validation Loss: 3.965389
Running corrects tensor(118, device='cuda:0')
Epoch: 22 	Training Loss: 1.839412 	Validation Loss: 4.076994
Running corrects tensor(131, device='cuda:0')
Epoch: 23 	Training Loss: 1.794916 	Validation Loss: 3.991258
Running corrects tensor(131, device='cuda:0')
Epoch: 24 	Training Loss: 1.722907 	Validation Loss: 4.063962
Running corrects tensor(130, device='cuda:0')
Epoch: 25 	Training Loss: 1.670196 	Validation Loss: 4.105696
Running corrects tensor(94, device='cuda:0')
Epoch: 26 	Training Loss: 1.702776 	Validation Loss: 4.578109
Running corrects tensor(138, device='cuda:0')
Epoch: 27 	Training Loss: 1.696352 	Validation Loss: 4.048858
Running corrects tensor(129, device='cuda:0')
Epoch: 28 	Training Loss: 1.669659 	Validation Loss: 4.046458
Running corrects tensor(95, device='cuda:0')
Epoch: 29 	Training Loss: 1.741965 	Validation Loss: 4.203661
Running corrects tensor(111, device='cuda:0')
Epoch: 30 	Training Loss: 1.720479 	Validation Loss: 4.358152
Running corrects tensor(117, device='cuda:0')
Epoch: 31 	Training Loss: 1.682501 	Validation Loss: 4.528480
Running corrects tensor(133, device='cuda:0')
Epoch: 32 	Training Loss: 1.713914 	Validation Loss: 4.109962
Running corrects tensor(107, device='cuda:0')
Epoch: 33 	Training Loss: 1.672913 	Validation Loss: 4.609086
Running corrects tensor(4, device='cuda:0')
Epoch: 34 	Training Loss: nan 	Validation Loss: nan
Running corrects tensor(4, device='cuda:0')
Epoch: 35 	Training Loss: nan 	Validation Loss: nan
Running corrects tensor(4, device='cuda:0')
Epoch: 36 	Training Loss: nan 	Validation Loss: nan
Running corrects tensor(4, device='cuda:0')
Epoch: 37 	Training Loss: nan 	Validation Loss: nan
Running corrects tensor(4, device='cuda:0')
Epoch: 38 	Training Loss: nan 	Validation Loss: nan
Running corrects tensor(4, device='cuda:0')
Epoch: 39 	Training Loss: nan 	Validation Loss: nan
Running corrects tensor(4, device='cuda:0')
Epoch: 40 	Training Loss: nan 	Validation Loss: nan
Running corrects tensor(4, device='cuda:0')
Epoch: 41 	Training Loss: nan 	Validation Loss: nan
Running corrects tensor(4, device='cuda:0')
Epoch: 42 	Training Loss: nan 	Validation Loss: nan
Running corrects tensor(4, device='cuda:0')
Epoch: 43 	Training Loss: nan 	Validation Loss: nan
Running corrects tensor(4, device='cuda:0')
Epoch: 44 	Training Loss: nan 	Validation Loss: nan
Running corrects tensor(4, device='cuda:0')
Epoch: 45 	Training Loss: nan 	Validation Loss: nan
Running corrects tensor(4, device='cuda:0')
Epoch: 46 	Training Loss: nan 	Validation Loss: nan
Running corrects tensor(4, device='cuda:0')
Epoch: 47 	Training Loss: nan 	Validation Loss: nan
Running corrects tensor(4, device='cuda:0')
Epoch: 48 	Training Loss: nan 	Validation Loss: nan
Running corrects tensor(4, device='cuda:0')
Epoch: 49 	Training Loss: nan 	Validation Loss: nan
Running corrects tensor(4, device='cuda:0')
Epoch: 50 	Training Loss: nan 	Validation Loss: nan
Running corrects tensor(4, device='cuda:0')
Epoch: 51 	Training Loss: nan 	Validation Loss: nan
Running corrects tensor(4, device='cuda:0')
Epoch: 52 	Training Loss: nan 	Validation Loss: nan
Running corrects tensor(4, device='cuda:0')
Epoch: 53 	Training Loss: nan 	Validation Loss: nan
Running corrects tensor(4, device='cuda:0')
Epoch: 54 	Training Loss: nan 	Validation Loss: nan
Running corrects tensor(4, device='cuda:0')
Epoch: 55 	Training Loss: nan 	Validation Loss: nan
Running corrects tensor(4, device='cuda:0')
Epoch: 56 	Training Loss: nan 	Validation Loss: nan
Running corrects tensor(4, device='cuda:0')
Epoch: 57 	Training Loss: nan 	Validation Loss: nan
Running corrects tensor(4, device='cuda:0')
Epoch: 58 	Training Loss: nan 	Validation Loss: nan
Running corrects tensor(4, device='cuda:0')
Epoch: 59 	Training Loss: nan 	Validation Loss: nan
Running corrects tensor(4, device='cuda:0')
Epoch: 60 	Training Loss: nan 	Validation Loss: nan
In [ ]:
 
In [27]:
print(len(loaders_scratch['train']))
print(len(loaders_scratch['valid']))
print(len(loaders_scratch['test']))
334
42
42

(IMPLEMENTATION) Test the Model

Try out your model on the test dataset of dog images. Use the code cell below to calculate and print the test loss and accuracy. Ensure that your test accuracy is greater than 10%.

In [28]:
def test(loaders, model, criterion, use_cuda):

    # monitor test loss and accuracy
    test_loss = 0.
    correct = 0.
    total = 0.

    model.eval()
    for batch_idx, (data, target) in enumerate(loaders['test']):
        # move to GPU
        if use_cuda:
            data, target = data.cuda(), target.cuda()
        # forward pass: compute predicted outputs by passing inputs to the model
        output = model(data)
        # calculate the loss
        loss = criterion(output, target)
        # update average test loss 
        test_loss = test_loss + ((1 / (batch_idx + 1)) * (loss.data - test_loss))
        # convert output probabilities to predicted class
        pred = output.data.max(1, keepdim=True)[1]
        # compare predictions to true label
        correct += np.sum(np.squeeze(pred.eq(target.data.view_as(pred))).cpu().numpy())
        total += data.size(0)
            
    print('Test Loss: {:.6f}\n'.format(test_loss))

    print('\nTest Accuracy: %2d%% (%2d/%2d)' % (
        100. * correct / total, correct, total))

# call test function    
test(loaders_scratch, model_scratch, criterion_scratch, use_cuda)
Test Loss: 3.599756


Test Accuracy: 16% (139/836)

Step 4: Create a CNN to Classify Dog Breeds (using Transfer Learning)

You will now use transfer learning to create a CNN that can identify dog breed from images. Your CNN must attain at least 60% accuracy on the test set.

(IMPLEMENTATION) Specify Data Loaders for the Dog Dataset

Use the code cell below to write three separate data loaders for the training, validation, and test datasets of dog images (located at dogImages/train, dogImages/valid, and dogImages/test, respectively).

If you like, you are welcome to use the same data loaders from the previous step, when you created a CNN from scratch.

In [47]:
## TODO: Specify data loaders
## TODO: Specify data loaders
from torchvision import datasets, models, transforms
import torchvision
import torch
import os
data_transforms = {
    'train': transforms.Compose([
        transforms.RandomResizedCrop(224),
        transforms.RandomHorizontalFlip(),
        transforms.ToTensor(),
        transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])
    ]),
    'valid': transforms.Compose([
        transforms.Resize(256),
        transforms.CenterCrop(224),
        transforms.ToTensor(),
        transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])
    ]),
    'test': transforms.Compose([
        transforms.Resize(256),
        transforms.CenterCrop(224),
        transforms.ToTensor(),
        transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])
    ]),
}
data_dir = '/data/dog_images/'
data_transfer = {x: datasets.ImageFolder(os.path.join(data_dir, x),
                                          data_transforms[x])
                  for x in ['train', 'valid','test']}
loaders_transfer = {x: torch.utils.data.DataLoader(image_datasets[x], batch_size=4,
                                             shuffle=True, num_workers=4)
              for x in ['train', 'valid','test']}

(IMPLEMENTATION) Model Architecture

Use transfer learning to create a CNN to classify dog breed. Use the code cell below, and save your initialized model as the variable model_transfer.

In [48]:
import torchvision.models as models
import torch.nn as nn


## TODO: Specify model architecture 
model_transfer = torchvision.models.resnet18(pretrained=True)
for param in model_transfer.parameters():
    param.requires_grad = False

num_ftrs = model_transfer.fc.in_features
model_transfer.fc = nn.Linear(num_ftrs, 133)
print(model_transfer)

if use_cuda:
    model_transfer = model_transfer.cuda()
ResNet(
  (conv1): Conv2d(3, 64, kernel_size=(7, 7), stride=(2, 2), padding=(3, 3), bias=False)
  (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (relu): ReLU(inplace)
  (maxpool): MaxPool2d(kernel_size=3, stride=2, padding=1, dilation=1, ceil_mode=False)
  (layer1): Sequential(
    (0): BasicBlock(
      (conv1): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace)
      (conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
    )
    (1): BasicBlock(
      (conv1): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace)
      (conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
    )
  )
  (layer2): Sequential(
    (0): BasicBlock(
      (conv1): Conv2d(64, 128, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace)
      (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (downsample): Sequential(
        (0): Conv2d(64, 128, kernel_size=(1, 1), stride=(2, 2), bias=False)
        (1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      )
    )
    (1): BasicBlock(
      (conv1): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace)
      (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
    )
  )
  (layer3): Sequential(
    (0): BasicBlock(
      (conv1): Conv2d(128, 256, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace)
      (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (downsample): Sequential(
        (0): Conv2d(128, 256, kernel_size=(1, 1), stride=(2, 2), bias=False)
        (1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      )
    )
    (1): BasicBlock(
      (conv1): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace)
      (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
    )
  )
  (layer4): Sequential(
    (0): BasicBlock(
      (conv1): Conv2d(256, 512, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace)
      (conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (downsample): Sequential(
        (0): Conv2d(256, 512, kernel_size=(1, 1), stride=(2, 2), bias=False)
        (1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      )
    )
    (1): BasicBlock(
      (conv1): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace)
      (conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
    )
  )
  (avgpool): AvgPool2d(kernel_size=7, stride=1, padding=0)
  (fc): Linear(in_features=512, out_features=133, bias=True)
)

Question 5: Outline the steps you took to get to your final CNN architecture and your reasoning at each step. Describe why you think the architecture is suitable for the current problem.

Answer: ResNet-18 is a convolutional neural network that is 18 layers deep. You can load a pretrained version of the network trained on more than a million images from the ImageNet database . The pretrained network can classify images into 1000 object categories, such as keyboard, mouse, pencil, and many animals. As a result, the network has learned rich feature representations for a wide range of images. The network has an image input size of 224-by-224, so i chosed this network will produce better accuracy

(IMPLEMENTATION) Specify Loss Function and Optimizer

Use the next code cell to specify a loss function and optimizer. Save the chosen loss function as criterion_transfer, and the optimizer as optimizer_transfer below.

In [33]:
criterion_transfer = nn.CrossEntropyLoss()
optimizer_transfer = torch.optim.SGD(model_transfer.fc.parameters(),lr=0.001)

(IMPLEMENTATION) Train and Validate the Model

Train and validate your model in the code cell below. Save the final model parameters at filepath 'model_transfer.pt'.

In [36]:
# train the model
# the following import is required for training to be robust to truncated images
from PIL import ImageFile
ImageFile.LOAD_TRUNCATED_IMAGES = True

def train(n_epochs, loaders, model, optimizer, criterion, use_cuda, save_path):
    """returns trained model"""
    # initialize tracker for minimum validation loss
    valid_loss_min = np.Inf 
    
    for epoch in range(1, n_epochs+1):
        # initialize variables to monitor training and validation loss
        train_loss = 0.0
        valid_loss = 0.0
        running_corrects=0.0
        ###################
        # train the model #
        ###################
        model.train()
        for batch_idx, (data, target) in enumerate(loaders['train']):
            # move to GPU
            if use_cuda:
                data, target = data.cuda(), target.cuda()
                optimizer.zero_grad()
            # forward pass
            output = model(data)
            # calculate batch loss
            loss = criterion(output, target)
            # backward pass
            loss.backward()
            # parameter update
            optimizer.step()
            # update training loss
            train_loss += loss.item() * data.size(0)
        ######################    
        # validate the model #
        ######################
        model.eval()
        for batch_idx, (data, target) in enumerate(loaders['valid']):
            # move to GPU
            if use_cuda:
                data, target = data.cuda(), target.cuda()
                #optimizer.zero_grad()
            ## update the average validation loss
            with torch.no_grad():
            
                output = model(data)
            # batch loss
            loss = criterion(output, target)
            # update validation loss
            valid_loss += loss.item()*data.size(0)
            _, preds = torch.max(output, 1)
            running_corrects += torch.sum(preds == target.data)
            
        # calculate average losses
        train_loss = train_loss/len(loaders['train'].dataset)
        valid_loss = valid_loss/len(loaders['valid'].dataset)
            
        # print training/validation statistics 
        print("Running corrects",running_corrects)
        print('Epoch: {} \tTraining Loss: {:.6f} \tValidation Loss: {:.6f}'.format(
            epoch, 
            train_loss,
            valid_loss
            ))
        
        ## TODO: save the model if validation loss has decreased
        if valid_loss <= valid_loss_min:
            print('Validation loss decreased ({:.6f} --> {:.6f}).  Saving model ...'.format(
            valid_loss_min,
            valid_loss))
            torch.save(model.state_dict(), save_path)
            valid_loss_min = valid_loss
            
    # return trained model
    return model


# train the model

# load the model that got the best validation accuracy

model_transfer = train(20, loaders_transfer, model_transfer, optimizer_transfer, criterion_transfer, use_cuda, 'model_transfer.pt')

# load the model that got the best validation accuracy (uncomment the line below)
model_transfer.load_state_dict(torch.load('model_transfer.pt'))
Running corrects tensor(134, device='cuda:0')
Epoch: 1 	Training Loss: 4.772898 	Validation Loss: 4.150718
Validation loss decreased (inf --> 4.150718).  Saving model ...
Running corrects tensor(271, device='cuda:0')
Epoch: 2 	Training Loss: 4.360699 	Validation Loss: 3.432190
Validation loss decreased (4.150718 --> 3.432190).  Saving model ...
Running corrects tensor(366, device='cuda:0')
Epoch: 3 	Training Loss: 4.015295 	Validation Loss: 2.925679
Validation loss decreased (3.432190 --> 2.925679).  Saving model ...
Running corrects tensor(403, device='cuda:0')
Epoch: 4 	Training Loss: 3.736321 	Validation Loss: 2.549100
Validation loss decreased (2.925679 --> 2.549100).  Saving model ...
Running corrects tensor(471, device='cuda:0')
Epoch: 5 	Training Loss: 3.510636 	Validation Loss: 2.139789
Validation loss decreased (2.549100 --> 2.139789).  Saving model ...
Running corrects tensor(499, device='cuda:0')
Epoch: 6 	Training Loss: 3.296725 	Validation Loss: 1.898346
Validation loss decreased (2.139789 --> 1.898346).  Saving model ...
Running corrects tensor(534, device='cuda:0')
Epoch: 7 	Training Loss: 3.113982 	Validation Loss: 1.793270
Validation loss decreased (1.898346 --> 1.793270).  Saving model ...
Running corrects tensor(546, device='cuda:0')
Epoch: 8 	Training Loss: 2.995977 	Validation Loss: 1.533712
Validation loss decreased (1.793270 --> 1.533712).  Saving model ...
Running corrects tensor(548, device='cuda:0')
Epoch: 9 	Training Loss: 2.859887 	Validation Loss: 1.427314
Validation loss decreased (1.533712 --> 1.427314).  Saving model ...
Running corrects tensor(565, device='cuda:0')
Epoch: 10 	Training Loss: 2.749347 	Validation Loss: 1.337891
Validation loss decreased (1.427314 --> 1.337891).  Saving model ...
Running corrects tensor(577, device='cuda:0')
Epoch: 11 	Training Loss: 2.663907 	Validation Loss: 1.278082
Validation loss decreased (1.337891 --> 1.278082).  Saving model ...
Running corrects tensor(586, device='cuda:0')
Epoch: 12 	Training Loss: 2.563143 	Validation Loss: 1.166236
Validation loss decreased (1.278082 --> 1.166236).  Saving model ...
Running corrects tensor(600, device='cuda:0')
Epoch: 13 	Training Loss: 2.505575 	Validation Loss: 1.103261
Validation loss decreased (1.166236 --> 1.103261).  Saving model ...
Running corrects tensor(615, device='cuda:0')
Epoch: 14 	Training Loss: 2.410356 	Validation Loss: 1.069676
Validation loss decreased (1.103261 --> 1.069676).  Saving model ...
Running corrects tensor(603, device='cuda:0')
Epoch: 15 	Training Loss: 2.361208 	Validation Loss: 1.048761
Validation loss decreased (1.069676 --> 1.048761).  Saving model ...
Running corrects tensor(617, device='cuda:0')
Epoch: 16 	Training Loss: 2.306194 	Validation Loss: 0.974350
Validation loss decreased (1.048761 --> 0.974350).  Saving model ...
Running corrects tensor(622, device='cuda:0')
Epoch: 17 	Training Loss: 2.271927 	Validation Loss: 0.927793
Validation loss decreased (0.974350 --> 0.927793).  Saving model ...
Running corrects tensor(624, device='cuda:0')
Epoch: 18 	Training Loss: 2.201904 	Validation Loss: 0.921140
Validation loss decreased (0.927793 --> 0.921140).  Saving model ...
Running corrects tensor(623, device='cuda:0')
Epoch: 19 	Training Loss: 2.199833 	Validation Loss: 0.875811
Validation loss decreased (0.921140 --> 0.875811).  Saving model ...
Running corrects tensor(636, device='cuda:0')
Epoch: 20 	Training Loss: 2.140401 	Validation Loss: 0.833105
Validation loss decreased (0.875811 --> 0.833105).  Saving model ...

(IMPLEMENTATION) Test the Model

Try out your model on the test dataset of dog images. Use the code cell below to calculate and print the test loss and accuracy. Ensure that your test accuracy is greater than 60%.

In [37]:
def test(loaders, model, criterion, use_cuda):

    # monitor test loss and accuracy
    test_loss = 0.
    correct = 0.
    total = 0.

    model.eval()
    for batch_idx, (data, target) in enumerate(loaders['test']):
        # move to GPU
        if use_cuda:
            data, target = data.cuda(), target.cuda()
        # forward pass: compute predicted outputs by passing inputs to the model
        output = model(data)
        # calculate the loss
        loss = criterion(output, target)
        # update average test loss 
        test_loss = test_loss + ((1 / (batch_idx + 1)) * (loss.data - test_loss))
        # convert output probabilities to predicted class
        pred = output.data.max(1, keepdim=True)[1]
        # compare predictions to true label
        correct += np.sum(np.squeeze(pred.eq(target.data.view_as(pred))).cpu().numpy())
        total += data.size(0)
            
    print('Test Loss: {:.6f}\n'.format(test_loss))

    print('\nTest Accuracy: %2d%% (%2d/%2d)' % (
        100. * correct / total, correct, total))

# call test function    
test(loaders_transfer, model_transfer, criterion_transfer, use_cuda)
Test Loss: 0.817147


Test Accuracy: 76% (643/836)

(IMPLEMENTATION) Predict Dog Breed with the Model

Write a function that takes an image path as input and returns the dog breed (Affenpinscher, Afghan hound, etc) that is predicted by your model.

In [49]:
class_names = [item[4:].replace("_", " ") for item in data_transfer['train'].classes]
print(class_names)
['Affenpinscher', 'Afghan hound', 'Airedale terrier', 'Akita', 'Alaskan malamute', 'American eskimo dog', 'American foxhound', 'American staffordshire terrier', 'American water spaniel', 'Anatolian shepherd dog', 'Australian cattle dog', 'Australian shepherd', 'Australian terrier', 'Basenji', 'Basset hound', 'Beagle', 'Bearded collie', 'Beauceron', 'Bedlington terrier', 'Belgian malinois', 'Belgian sheepdog', 'Belgian tervuren', 'Bernese mountain dog', 'Bichon frise', 'Black and tan coonhound', 'Black russian terrier', 'Bloodhound', 'Bluetick coonhound', 'Border collie', 'Border terrier', 'Borzoi', 'Boston terrier', 'Bouvier des flandres', 'Boxer', 'Boykin spaniel', 'Briard', 'Brittany', 'Brussels griffon', 'Bull terrier', 'Bulldog', 'Bullmastiff', 'Cairn terrier', 'Canaan dog', 'Cane corso', 'Cardigan welsh corgi', 'Cavalier king charles spaniel', 'Chesapeake bay retriever', 'Chihuahua', 'Chinese crested', 'Chinese shar-pei', 'Chow chow', 'Clumber spaniel', 'Cocker spaniel', 'Collie', 'Curly-coated retriever', 'Dachshund', 'Dalmatian', 'Dandie dinmont terrier', 'Doberman pinscher', 'Dogue de bordeaux', 'English cocker spaniel', 'English setter', 'English springer spaniel', 'English toy spaniel', 'Entlebucher mountain dog', 'Field spaniel', 'Finnish spitz', 'Flat-coated retriever', 'French bulldog', 'German pinscher', 'German shepherd dog', 'German shorthaired pointer', 'German wirehaired pointer', 'Giant schnauzer', 'Glen of imaal terrier', 'Golden retriever', 'Gordon setter', 'Great dane', 'Great pyrenees', 'Greater swiss mountain dog', 'Greyhound', 'Havanese', 'Ibizan hound', 'Icelandic sheepdog', 'Irish red and white setter', 'Irish setter', 'Irish terrier', 'Irish water spaniel', 'Irish wolfhound', 'Italian greyhound', 'Japanese chin', 'Keeshond', 'Kerry blue terrier', 'Komondor', 'Kuvasz', 'Labrador retriever', 'Lakeland terrier', 'Leonberger', 'Lhasa apso', 'Lowchen', 'Maltese', 'Manchester terrier', 'Mastiff', 'Miniature schnauzer', 'Neapolitan mastiff', 'Newfoundland', 'Norfolk terrier', 'Norwegian buhund', 'Norwegian elkhound', 'Norwegian lundehund', 'Norwich terrier', 'Nova scotia duck tolling retriever', 'Old english sheepdog', 'Otterhound', 'Papillon', 'Parson russell terrier', 'Pekingese', 'Pembroke welsh corgi', 'Petit basset griffon vendeen', 'Pharaoh hound', 'Plott', 'Pointer', 'Pomeranian', 'Poodle', 'Portuguese water dog', 'Saint bernard', 'Silky terrier', 'Smooth fox terrier', 'Tibetan mastiff', 'Welsh springer spaniel', 'Wirehaired pointing griffon', 'Xoloitzcuintli', 'Yorkshire terrier']
In [53]:
### TODO: Write a function that takes a path to an image as input
### and returns the dog breed that is predicted by the model.

# list of class names by index, i.e. a name can be accessed like class_names[0]
class_names = [item[4:].replace("_", " ") for item in data_transfer['train'].classes]

def predict_breed_transfer(img_path):
    transform = transforms.Compose([transforms.Resize(256),transforms.CenterCrop(224),transforms.ToTensor(),transforms.Normalize(mean=[0.485, 0.456, 0.406],std=[0.229, 0.224, 0.225])])
    img = Image.open(img_path)
    transform_image = transform(img)
    squ = torch.unsqueeze(transform_image, 0)
    #img = Variable(squ)
    model_transfer.eval()
    ## Load and pre-process an image from the given img_path
    ## Return the *index* of the predicted class for that image
    prediction = model_transfer(squ.cuda())  # Returns a Tensor of shape (batch, num class labels)
    prediction = prediction.cpu().data.numpy().argmax()
    name = class_names[prediction]
    return name
    
    
    
In [55]:
dog_files = np.array(glob("/data/dog_images/*/*/*"))
dog_files_short = dog_files[:10]
for i in range(len(dog_files_short)):
    result = predict_breed_transfer(dog_files_short[i])
    print("{} result {}".format(i,result))
0 result Keeshond
1 result Komondor
2 result Keeshond
3 result Black and tan coonhound
4 result Border collie
5 result Black and tan coonhound
6 result Norwegian lundehund
7 result Border collie
8 result Border collie
9 result Keeshond

Step 5: Write your Algorithm

Write an algorithm that accepts a file path to an image and first determines whether the image contains a human, dog, or neither. Then,

  • if a dog is detected in the image, return the predicted breed.
  • if a human is detected in the image, return the resembling dog breed.
  • if neither is detected in the image, provide output that indicates an error.

You are welcome to write your own functions for detecting humans and dogs in images, but feel free to use the face_detector and human_detector functions developed above. You are required to use your CNN from Step 4 to predict dog breed.

Some sample output for our algorithm is provided below, but feel free to design your own user experience!

Sample Human Output

(IMPLEMENTATION) Write your Algorithm

In [73]:
### TODO: Write your algorithm.
### Feel free to use as many code cells as needed.
import matplotlib.pyplot as plt
import cv2 as cv
def imshow(inp, title=None):
    """Imshow for Tensor."""
    inp = inp.numpy().transpose((1, 2, 0))
    mean = np.array([0.485, 0.456, 0.406])
    std = np.array([0.229, 0.224, 0.225])
    inp = std * inp + mean
    inp = np.clip(inp, 0, 1)
    plt.imshow(inp)
    if title is not None:
        plt.title(title)
    plt.pause(0.001) 

def run_app(img_path):
    img = Image.open(img_path)
    plt.imshow(img)
    
    
    detect_human_faces = face_detector(img_path)
    if detect_human_faces == True:
        print("Hello,human")
        plt.show()
        predict_dog =  predict_breed_transfer(img_path)
        
        print('You look like a', predict_dog)
    elif (dog_detector(img_path))==True:
        print("Hello, Dog")
        plt.show()
        predict_dog = predict_breed_transfer(img_path)
        
        print('You look like a',predict_dog)
    else:
        
        print("I am not able recognize")
        
        
        
    
    
    
    ## handle cases for a human face, dog, and neither
    
In [ ]:
 

Step 6: Test Your Algorithm

In this section, you will take your new algorithm for a spin! What kind of dog does the algorithm think that you look like? If you have a dog, does it predict your dog's breed accurately? If you have a cat, does it mistakenly think that your cat is a dog?

(IMPLEMENTATION) Test Your Algorithm on Sample Images!

Test your algorithm at least six images on your computer. Feel free to use any images you like. Use at least two human and two dog images.

Question 6: Is the output better than you expected :) ? Or worse :( ? Provide at least three possible points of improvement for your algorithm.

Answer: Output is better first, i like would to add more data augment for this project to get even better accuracy Second, training datasets is less, so if dataset have fews images more about on breed, i will be good predictor third, in training images contain other features with dog images

In [74]:
## TODO: Execute your algorithm from Step 6 on
## at least 6 images on your computer.
## Feel free to use as many code cells as needed.

## suggested code, below
for file in np.hstack((human_files[:10], dog_files[:10])):
    run_app(file)
Hello,human
You look like a Black and tan coonhound
Hello,human
You look like a Black and tan coonhound
Hello,human
You look like a Dachshund
Hello,human
You look like a Norwegian lundehund
Hello,human
You look like a Akita
Hello,human
You look like a Dachshund
Hello,human
You look like a Brussels griffon
Hello,human
You look like a Black and tan coonhound
Hello,human
You look like a Brussels griffon
Hello,human
You look like a Entlebucher mountain dog
Hello, Dog
You look like a Keeshond
Hello, Dog
You look like a Komondor
Hello, Dog
You look like a Keeshond
Hello, Dog
You look like a Black and tan coonhound
Hello, Dog
You look like a Border collie
Hello, Dog
You look like a Black and tan coonhound
Hello, Dog
You look like a Norwegian lundehund
Hello, Dog
You look like a Border collie
Hello, Dog
You look like a Border collie
Hello, Dog
You look like a Keeshond
In [ ]: